A Fuzzy-set-based Joint Distribution Adaptation Method for Regression and its Application to Online Damage Quantification for Structural Digital Twin. (arXiv:2211.02656v1 [cs.CE])

Online damage quantification suffers from insufficient labeled data. In this
context, adopting the domain adaptation on historical labeled data from similar
structures/damages to assist the current diagnosis task would be beneficial.
However, most domain adaptation methods are designed for classification and
cannot efficiently address damage quantification, a regression problem with
continuous real-valued labels. This study first proposes a novel domain
adaptation method, the Online Fuzzy-set-based Joint Distribution Adaptation for
Regression, to address this challenge. By converting the continuous real-valued
labels to fuzzy class labels via fuzzy sets, the conditional distribution
discrepancy is measured, and domain adaptation can simultaneously consider the
marginal and conditional distribution for the regression task. Furthermore, a
framework of online damage quantification integrated with the proposed domain
adaptation method is presented. The method has been verified with an example of
a damaged helicopter panel, in which domain adaptations are conducted across
different damage locations and from simulation to experiment, proving the
accuracy of damage quantification can be improved significantly even in a noisy
environment. It is expected that the proposed approach to be applied to the
fleet-level digital twin considering the individual differences.